Professional Writing Service http://StudyHub.vip/An-Ontological-And-Constructional-Appro 👈

1. 63
Chapter 3
An ontological and constructional approach to the discourse analysis of
commemorative speeches in Croatia
Benedikt Perak
0000-0003-4177-5307
Abstract
This chapter discusses the methods and results of an ontological, conceptual and linguistic analysis
of collective identities and the sociocultural concepts in staged communication during
commemoration rituals. The lexical data for the study is provided by a corpus of sixty-one speeches
delivered at seven commemoration sites from 2014-2016, comprising 56,291 tokens. By using the
graph theory algorithms on the level of lexical concepts we classified sixty-four speakers and
eighteen supporting institutions according to the 3,370 invoked noun concepts at the
commemorations. The classification process has revealed distinct communities of speakers and
their shared choice of salient concepts and strategies of framing the affective dispositions and
cognitive processes that form the basis for the construction of group identities, interaction and
communication practices, political agendas and the dominant cultural model of national identity
in general.
Introduction
This chapter 1
deals with the commemoration rituals as communication practices and
conceptualization mechanisms. Particularly, the study analyses commemoration speeches
delivered at the seven commemoration sites monitored by the FRAMNAT project from January
2014 to December 2016. The transcription of the speeches enabled the creation of the FRAMNAT
2014-2016 corpus and the cultural cognitive discourse analysis of the texts. The speeches are seen
as a network of conceptualizations about the referential historical events in Croatian cultural
memory, construed with the function to reinforce a range of bio-psycho-social phenomena in the
commemoration participants. The corpus analysis measured the frequency of the activated
concepts in speeches by speakers and institutions. By using the graph theory algorithms on the
level of lexical concepts we classified sixty-four speakers and eighteen supporting institutions
1
The author acknowledges full support of the FRAMNAT project HRZZ 3782, funded by the Croatian Science
Foundation.

2. 64
according to the 3,370 invoked noun concepts at the commemorations. The classification process
has revealed distinct communities of speakers and their shared choice of salient concepts and
strategies of framing the affective dispositions and cognitive processes that form the basis for the
construction of group identities, interaction and communication practices, political agendas and
the dominant cultural model of national identity in general.
Speakers at the commemorations
The central role of the speakers in the commemoration is to conceptualize the referential traumatic
event in history by captivating the attention and raising the motivation of the listeners, providing
reasoning and establishing culturally normative values (Charteris-Black, 2005; 2006; Pavlaković
and Perak, 2017). The speeches are typically performed by a single speaker and addressed
primarily to the assembled audience at the commemoration site, and secondarily to the wider
national audience through media coverage. Each speaker is connected and supported by some
institution that partakes in the political agenda of the commemoration.
The data in this chapter presents structure of the sixty-four speakers at the seven commemoration
sites from 2014-2016. The list of speakers ordered by the number of speeches is the following;
Franjo Habulin, president of the Association of Antifascist Fighters and Antifascists of Croatia
and former Prime Minister Zoran Milanović each produced five speeches. Milorad Pupovac,
representative of the Serbian minority in the Croatian Parliament, delivered four speeches.
Speakers with three speeches in the corpus are President Kolinda Grabar-Kitarović, former
president Stjepan Mesić, former Parliament Speaker Josip Leko, former President Ivo Josipović,
and Sisak Mayor Kristina Ikić Baniček. Speakers with two speeches are former Parliament Speaker
Željko Reiner, Nataša Mataušić, Representative Bruna Esih, Zoran Pusić, Cardinal Josip Bozanić,
bishop Mate Uzinić, archbishop Đuro Hranić, Ante Kutleša, Milan Tankosić, Idriz ef. Bešić,
military bishop Juraj Jezerinac, Milan Surla, and Dragan Čović. Speakers with one speech are:
Zagreb mayor Milan Bandić, bishop Nikola Kekić, Aziz ef. Hasanović, Borjana Krišto, Nevenka
Marinković, Orest Wilczynski, Frano Čirko, Maciej Szymanski, archbishop Želimir Puljić, bishop
Franjo Komarica, Anneliese Kitzmüller, Ivica Jagodić, Nikola Budija, Ivanka Roksandić, Zlatko
Ževrnja, historian Dragan Markovina, Knin mayor Josipa Rimac, Branko Lustig, Dražimir Jukić,
Dubravka Jurlina Alibegović, Manda Patko, Ivica Vukelić, bishop Vjekoslav Huzjak, Tomislav
Sopta, Imam Admir Muhić, Boris Prebeg, Ivica Glavota, Ivan Vukić, Margareta Mađerić, Madij
Ismailov, and several anonymous students from the Elementary School “Blago Zadro” in Borovo
naselje.2
2
Biographical details of the key speakers are given in the subsequent chapters of this volume.

3. 65
The structure of speakers at the commemorations is represented in Illustration 1.
Figure 3.1: Graphical representation of the speakers at the commemorations. The size of the nodes
is represented relative to the amount of connections with other nodes (degree).
The layout of the graph is produced by connecting a speaker to the commemoration site where the
speech was delivered. The majority of the speakers have delivered speeches at only one
commemoration site, but some of them, mostly high ranking political officials, have appeared in
several commemorations, such as the former president Ivo Josipović, who delivered speeches in
Knin, Brezovica, and Jasenovac, as did former Prime Minister Zoran Milanović. Kolinda Grabar-
Kitarović, elected president in January 2015, appeared as a speaker in Knin and Brezovica.
Cardinal Josip Bozanić and other members of the Catholic Church also appeared at several
commemorations including Knin, Vukovar, and Bleiburg.
The network representation of the graph illustrates the politically polarized structure of the
commemorations with Srb, Brezovica, and Jasenovac on the left side and Jazovka and Bleiburg on
the other. As elaborated in other chapters, the commemorations in Srb, Brezovica, and Jasenovac
promote an antifascist cultural memory, while the Jazovka and Bleiburg sites commemorate crimes
committed by the same army, Tito’s Partisans, which fought for antifascist values. Speakers in all
of these commemorations act as the memory agents of the traumatic past of the Second World War
in Croatia. However, given the fact that these speakers often represent a political or clerical
institution and support its worldview, this graph presents interesting information about the
structure of the promoted cultural memory as well as the influence of the political agenda and

4. 66
ideology that, consequently, contributes to the contemporary social distribution of the cultural
representation and conceptual framing of the Croatian identity.
Two commemorations related to the Croatian War of Independence (Homeland War), Vukovar
and Knin, are different with respect to the communication structure. The Vukovar commemoration
is distinctive for its lack of overt verbal public messages by the political officials at the site. Instead,
the commemoration includes a Procession of Memory from the Vukovar hospital to the Homeland
War Memorial Cemetery where the political representatives, state and local officials, war veterans
and victims’ organizations lay wreaths (see Chapters 8 and 9). The subsequent speeches on the
memorial site in Vukovar are held exclusively by clerical representatives of the Catholic Church.
Along with promoting Christian theological and liturgical values, these speeches have a socio-
political function that conflates the heightened emotional remembrance of the Vukovar victims
with the Christian ontological belief in soul and afterlife while framing contemporary Croatian
identity as Catholic denomination.
On the other hand, the commemoration in Knin is framed in terms of Victory and Homeland
Thanksgiving Day and the Day of Croatian Defenders, celebrating the reintegration of Croatian
territory, and therefore an unavoidable place for high political representatives, representatives of
the government, and veteran’s organizations. It is interesting to note that from 2015, high
representatives of the Catholic Church have shared this prominent public communication space,
besides the usual church service and organized Mass. This practice is obviously correlated with
the political rise of the HDZ party in the Parliament and the party’s victory in the presidential
election in the beginning of 2015. The structure of the speakers and their supporting institutions in
Knin from 2014 to 2016 is represented in Illustration 2.

5. 67
Figure 3.2 The speakers and the supporting institutions at the Knin commemoration from 2014-
2016.
It can be argued that the changing structure of the speakers and supporting institutions in the Knin
commemoration reflects the dynamic of the political power and the coinciding surge of the
conservative religious-political movement in Croatia (Petričušić, Čehulić, Čepo 2017).
Commemorative speeches as texts
From 2014 to the end of 2016, the above-mentioned speakers delivered a total of 101
commemorative speeches.
Table 3.1 Number of speeches per commemoration.
Commemoration Number of speeches
Brezovica 17
Bleiburg 17
Srb 15

6. 68
Knin 12
Jasenovac 12
Vukovar - Borovo selo 11
Jazovka 8
Vukovar 7
As a part of the FRAMNAT methodology, the texts of the speeches were transcribed and stored in
an online spreadsheet 3
with relevant meta-data. Using a Python programming language
(https://www.python.org/), Py2Neo library (http://py2neo.org/v3/) and a Neo4j property graph
database (https://neo4j.com/) the data was converted into a graph property data model. The texts
were further tokenized, lemmatized, and parsed using the Reldi API Parser library (Ljubešić et al.,
2016) with respective number of tokens, morphosyntactic forms of tokens, part of speech, number
of words, number of lemmas and dependency functions stored as properties of the instances of the
classes (labels) according to the relation model represented in the illustration 3.
3
https://docs.google.com/spreadsheets/d/1rXV9x9-Jdpw84nmcOTEJBHnd-S5nu7-
YDYk8zj06sN8/edit?usp=sharing

7. 69
Figure 3.3 Schema of the graph property database model with nodes as the data classes and
relationships as ontological connections between classes.
Table 3.2 Number of tokens, words, lemmas and sentences in the FRAMNAT corpus 2014-2016.
Class Number
All Tokens (including punctuation signs) 80236
Word tokens 71006
Words 16727
Lemma 7687
Sentences 3314
The graph of the property data model (Illustration 3) allows us to create specific queries about the
structure of relations between instances of the interconnected classes (ontological entities). To
begin with, we can search for the speakers with the lengthiest speeches, see where and when were
they delivered (Table 3), or we can get the statistical average and standard deviation about the
number of sentences delivered by a speaker (Table 4).
Table 3.3 List of speeches with the highest number of sentences and their corresponding speakers.
Author Number of sentences Commemoration Year
1 Vjekoslav Huzjak 175 Jazovka 2016
2 Nikola Kekić 105 Jazovka 2014
3 Josip Bozanić 93 Bleiburg 2015
4 Franjo Komarica 93 Bleiburg 2016
5 Josip Bozanić 90 Knin 2015
6 Mate Uzinić 87 Vukovar 2014
7 Želimir Puljić 77 Vukovar 2015
8 Ivica Glavota 74 Knin 2016

8. 70
9 Kolinda Grabar-
Kitarović
72 Knin 2016
10 Stjepan Mesić 67 Srb 2014
11 Kolinda Grabar-
Kitarović
62 Knin 2015
12 Milorad Pupovac 61 Srb 2014
13 Ivo Josipović 57 Brezovica 2014
14 Milinko Čekić 54 Jasenovac 2014
15 Mate Uzinić 53 Bleiburg 2014
16 Milorad Pupovac 52 Srb 2016
17 Zoran Milanović 52 Brezovica 2014
18 Zoran Milanović 50 Brezovica 2015
19 Zoran Milanović 50 Jasenovac 2014
Table 3.4 Average number of sentences per speaker
Speaker Average number of
sentences per speaker
Standard deviation
1 Vjekoslav Huzjak 175 0
2 Nikola Kekić 105 0
3 Franjo Komarica 93 0
4 Josip Bozanić 91.5 2.1213
5 Želimir Puljić 77 0
6 Ivica Glavota 74 0
7 Mate Uzinić 70 24.041
8 Kolinda Grabar-Kitarović 58.6667 15.275
9 Milinko Čekić 54 0
10 Stjepan Mesić 51.3333 13.796

9. 71
11 Boris Prebeg 46 0
12 Milan Bandić 45 0
13 Milorad Pupovac 43.75 16.214
14 Dražimir Jukić 41 0
15 Tomislav Sopta 41 0
16 Zoran Milanović 39.8 15.006
17 Bruna Esih 39.5 4.9497
18 Ivo Josipović 39.3333 22.501
…
The role and influence of memory agents often has to do not only with the aptitude and eloquence
of the speaker, but with the institution they represent. From Tables 3 and 4 we can note that the
longest speeches were delivered by the representatives of the Catholic Church, which can also be
seen on Tables 5 and 6. The representatives of the Catholic Church in commemorations have
produced at least 40 percent more sentences than speakers from any other political institution,
Second World War veterans, or Homeland War veteran associations, with an average of 61.4
sentences per speech and a standard deviation of 49.7 sentences.
Table 3.5 Sum of the sentences delivered by the representatives of an institution
Institution Sentences
Catholic Church in Croatia 675
President of the Republic of Croatia 373
Association of Anti-Fascist Fighters and Anti-Fascists of the Republic of
Croatia
311
Croatian Government 243
Serb National Council 175
Elementary School “Blago Zadro”, Borovo naselje 143
Parliament of the Republic of Croatia 143
The Greek Catholic Church in Croatia 105

10. 72
The Catholic Church in Bosnia 93
The Islamic Community of Croatia 87
Honorary Bleiburg Guard 82
Hrvatski Obredni Zdrug Jazovka 75
Defenders of the Homeland War 74
City of Sisak 54
Public institution of the Jasenovac Memorial Area 51
The Government of the Federation of Bosnia and Herzegovina 51
Antifascist League of the Republic of Croatia 48
City of Zagreb 45
Association of Veterans of Croatian Guardian Units 41
Municipality of Gračac 37
County of Split-Dalmatia 36
The Embassy of Poland 33
Military Ordinate in the Republic of Croatia 31
Association of Antifascist Fighters and Antifascists of the city of Zadar 31
Association of the 6th
Lika Division 28
The Armed Forces of the Republic of Croatia 27
City of Knin 23
Vukovar mothers 23
County of Sisak-Moslavina 21
HDZ BiH 20
Greek Catholic Church in Ukraine 12
Diplomatic Corps of the Antifascist Coalition countries 10
Table 3.6 Average number of sentences per speech delivered by the representative of an Institution

11. 73
Institution Average SD
Greek Catholic Church in Croatia 105 0
the Catholic Church in Bosnia 93 0
Defenders of the Homeland War 74 0
the Catholic Church in Croatia 61.3636 49.7740
President of the Republic of Croatia 46.625 17.7276
City of Zagreb 45 0
Serb National Council 43.75 16.2147
Association of Veterans of Croatian Guardian Units 41 0
Association of Anti-Fascist Fighters and Anti-Fascists of the
Republic of Croatia
38.875 14.8847
County of Split-Dalmatia 36 0
Croatian Government 34.7143 16.3066
the Embassy of Poland 33 0
Association of Anti-Fascist Fighters and Anti-Fascists of the City of
Zadar
31 0
Honorary Bleiburg Guard 27.3333 11.8462
the Armed Forces of the Republic of Croatia 27 0
Public institution of the Jasenovac Memorial Area 25.5 4.94975
the Government of the Federation of Bosnia and Herzegovina 25.5 16.2635
Hrvatski Obredni Zdrug Jazovka 25 19
Anti-fascist League of the Republic of Croatia 24 7.07107
parliament of the Republic of Croatia 23.8333 7.02614
…
Using the combined measures of the absolute length and average length of speeches, we can
conclude that other significant institutions include the President of the Republic of Croatia, the

12. 74
Association of Anti-Fascist Fighters and Anti-Fascists of the Republic of Croatia, the Croatian
Government, and the Serb National Council.
In terms of the formation of social ontology (Searle, 2010), it is argued that the speech length
feature is an indication of the institution’s cultural dissemination power and political influence in
commemoration practices. The dissemination enables the intersubjective sharing of a set of
cognitive schemas within a social group, defined as a Cultural Model (D’Andrade, 1987: 112).
According to this feature, the data in Tables 5 and 6 indicate the strong cultural hegemony of the
Catholic Cultural Model in the framing of the collective identity in the commemoration speeches.
The proliferation of the Catholic model influences other political models and is sometimes in
opposition with other cultural models. The political influence is subsequently associated with the
in-group identification and recipient’s approval of the Cultural Model profiled by the
representatives in their speeches.
A good example of these social dynamics in commemorative events is the speech by Prime
Minister Milanović delivered in Knin 2014. Milanović’s speeches in the FRAMNAT corpus
usually have 40 sentences with standard deviation of 15 sentences (Table 4). However, at that
commemoration in Knin he conveyed only 21 sentences.4
The speech was constantly interrupted
by a rather large group of right-wing nationalists who were relentlessly disapproving every word,
and moreover, some of them started to sing traditional Croatian patriotic songs during his speech.
This was not so much caused by the disapproval of a speech itself, rather it was an overt political
denunciation of a different cultural model represented by Milanović himself, the president of a
left-wing political party (SDP – Social Democratic Party) and prime minister of a center-left
coalition. Due to the inability to establish a prototypical speaker-listener relation he had to shorten
the speech. The length of a speech can thus represent the hegemonic acceptance/disapproval of the
institutional deontic power (Searle, 2010), cultural frames, along with the reinforcement or
opposition of the group identity (Ma´ iz, 2003; Hogan, 2009; Pavlaković and Perak, 2017).
Embodied cognition and the ontological model of the texts
The texts are conceptually analyzed from the perspective of embodied cognition theory and
compatible methodologies. Embodied cognition approaches to communication (Bergen et al.,
2004, Lakoff, 1987; 2008; Lakoff & Johnson, 1999) argue that the understanding of a linguistic
expression involves a mental simulation and/or enactment of the appropriate embodied experience.
For instance, the processing of a sentence, such as in example 1 below, involves the syntactic
processing of tokens that is represented with the ‘NEXT_token’ sequence in Illustration 4. Each
token is recognized as a word with a set of morphosyntactic features that are conceptually mapped
onto a lexical concept (lemma). The embodied perspective argues that the meaning of a sentence
emerges from the neuro-cognitive recreation of the superimposed mental simulations construed by
the syntactic and semantic features of the symbolically activated concepts. The processing of a
4
https://www.youtube.com/watch?v=ymLuchbuSC0

13. 75
lexical concept involves the mental simulation of the referential experience as well as the
processing of the syntactic and semantic properties of the lexeme. For example, the adjective
modifier teška (“heavy”) involves the mental simulation of heaviness, while krvava (“bloody”)
instantiates the recreation of the injury, blood, and physical harm. These adjectives profile the
noun borba (“fight”) that is the direct object of a process početi (“begin”) and is a noun modifier
of a noun phrase konačna sloboda (“final freedom”), related with a preposition za (“for”) that
logically specifies the purpose of the violent and painful simulation of the fighting.
1) Počela je teška i krvava borba za konačnu slobodu. (Kekić, Jazovka, 2014)
“A heavy and bloody fight for final freedom has begun.”
The dynamic cognitive process of meaning creation profiles a referential reality by activating the
embodied experience of the conceptualizer. The communicative act of conceptualization thus
frames the neuro-psychological states of the listener and influences their inferential configuration
and behavioral outcomes.
Figure 3.4 The syntactic, morphosyntactic, and conceptual relation in a sentence of a text delivered
by a speaker.
The data model (Illustrations 3 and 4), therefore, connects each Text with the class Tokens, and
schematically maps the morphosyntactic properties of individual tokens and their grammatical
relations to the instances of the words and lemmas classes. The lemma, the basic linguistic form
of a word, is schematized as the concept expressed in a language code.
Frequency as a measure of the cognitive focus
The ontological model allows for the creation of queries on sequential, syntactic, and conceptual
three levels of abstraction. By analyzing the frequency of the lemmas we can reveal the saliency
of the conceptual entities in the FRAMNAT corpora. On the conceptual level, the frequency

14. 76
expresses the intention of the speakers to focus the attention of the listeners to a specific
phenomenon. In this study we will present only the analysis of the noun lexical concepts. The ten
most frequent nouns concepts in the whole corpus are represented in Table 7.
Table 3.7 The list of ten most frequent noun concepts in the FRAMNAT 2014-2016 corpus.
Lemma Frequency
1 Hrvatska “Croatia” 486
2 narod “people” 323
3 godina “year” 322
4 čovjek “man” 308
5 žrtva “victim” 269
6 dan “day” 226
7 rat “war” 219
8 život “life” 195
9 država “state” 188
10 istina “truth” 182
11 mjesto “place” 174
12 zločin “crime” 157
13 sloboda “freedom” 154
14 borba “struggle” 150
15 domovina “homeland” 144
16 branitelj “defender” 137
17 put “path” 135
18 zlo “evil” 122
19 grad “city” 121
20 povijest “history” 121
…

15. 77
One of the interesting concepts in this frequency list is the word domovina “homeland”. The word
etymologically refers to the concept home (Latin domus, Old Church Slavic domъ), extending the
home feeling to the land, or even metonymically and metaphorically to the state. Table 8 lists
speeches with more than five occurrences of this lemma.
Table 3.8 List of texts with five or more occurrences of lemma domovina (“homeland”).
Commemoration Year Speaker Frequency
Jazovka 2016 Vjekoslav Huzjak 30
Vukovar 2014 Mate Uzinić 9
Knin 2015 Josip Bozanić 9
Knin 2015 Kolinda Grabar-Kitarović 7
Srb 2016 Nikola Budija 6
Knin 2016 Juraj Jezerinac 6
Bleiburg 2014 Mate Uzinić 6
Knin 2016 Kolinda Grabar-Kitarović 5
Jazovka 2014 Boris Prebeg 5
Jazovka 2016 Frano Čirko 5
The word domovina (“homeland”) is used most frequently in speeches delivered at Jazovka,
Vukovar, Knin, and in a speech in Srb and Bleiburg. The usage is indicative of the
commemorations with dominantly national patriotic sentiment. To see whether this presumption
is accurate we checked the frequency of the usage in terms of the institutions (Table 9).
Table 3.9 The frequency of the word domovina (“homeland”) per institution
Institution Frequency Average Standard
Deviation
Texts
1 Catholic Church in Croatia 73 7.300 8.4070 10
2 President of the Republic of Croatia 18 2.571 2.4398 7
3 Hrvatski Obredni Zdrug Jazovka 10 5 0 2
4 Elementary School “Blago Zadro”, Borovo
naselje
9 2.25 0.9574 4

16. 78
5 Association of Anti-Fascist Fighters and
Anti-Fascists of the City of Zadar
6 6 0 1
6 Military Ordinary in the Republic of
Croatia
6 6 0 1
7 Vukovar mothers 3 3 0 1
8 City of Knin 3 3 0 1
9 Parliament of the Republic of Croatia 3 1.5 0.7071 2
10 Honorary Bleiburg Platoon 3 1.5 0.7071 2
11 Association of Veterans of Croatian
Guardian Units
3 3 0 1
12 Defenders of the Homeland War 2 2 0 1
13 Antifascist League of the Republic of
Croatia
2 2 0 1
13 Greek Catholic Church in Croatia 2 2 0 1
15 The Government of the Federation of
Bosnia and Herzegovina
1 1 0 1
16 County of Sisak-Moslavina 1 1 0 1
17 The Catholic Church in Bosnia 1 1 0 1
18 The Armed Forces of the Republic of
Croatia
1 1 0 1
The usage of the word homeland is most frequently related to the representatives of the Catholic
Church, the institution of the President, veterans of the Homeland War, and the antifascists from
Zadar. By comparing the list of institutions with most sentences in the commemoration (Table 5)
we can note that representatives of the Serb National Council and the Islamic Community of
Croatia did not instantiate the conceptualization of the word domovina (“homeland”). Does this
mean that they have different types of conceptualization models of the state? This question can be
analyzed with the ontological corpus analysis (OCA). OCA opens the possibilities of the empirical
approach to investigate the conceptual variation of the promoted Cultural Models between
institutions by measuring the preference of the used words. The hypothesis is that the difference
between the configurations of a cultural model represents contesting conceptualizations within a
culture system induced by specific intra-cultural perspectives.

17. 79
Identifying the Institutional distribution of concepts
By formulating a query that extracts the frequency of the lemmatized nouns used by the
representatives of an institution in their speeches we produced a graph that illustrates the concepts
that are common to many institutions, and vice versa, concepts that are specific to a certain
institution (Illustration 5).
Figure 3.5 Graph of the relations between the 3,370 noun lemmas expressed by the representatives
of 31 institutions. The size of the labels and nodes corresponds to the overall frequency of the
nouns connected with a particular Institution. The graph is projected in three ordinates: x, y and z.
The projected z ordinate, perceived as the height of the institution nodes, corresponds to the
number of different words connected in the graph.
It is not convenient to reproduce the whole graph in this print edition due to the spatial restrictions
of representing 3,401 nodes (3,370 nouns and thirty-one institutions) and 8,492 relations.

18. 80
However, the graph can be interactively explored on the FRAMNAT web site.5
It is important to
notice the connectedness and the structure between the institutions and concepts, represented by
the Force Layout with the z ordinate in Illustration 5. The nouns commonly used by many
representatives of the institutions are located in the oval center of the graph due to the many
connections with different representatives, while the nouns specific and unique to a certain
institution extend to the margins.
Nominal concepts used by the representatives of all institutions, with 31 degrees, are Hrvatska
(“Croatia”) and godina (“year”). These words are semantically necessary and therefore not
distinctive in terms of the specific intra-cultural conceptualization analysis. On the other hand,
there are 1,941 words (53%) with degree 1, specific to an institution and their Cultural Model.
Although we cannot argue that every concept with degree 1 expresses some specific feature of the
Cultural Model, they obviously contribute to the uniqueness of the conceptualization strategy. For
instance, Milorad Pupovac, representative of the Serb National Council at the 2015 antifascist
commemoration in Srb, addressed the gathered participants (see example 2) while a right-wing
counter commemoration was ongoing only a few hundred meters away. The two commemoration
groups were separated by strong police forces.
2) …država štiti nas ovdje kao da smo u rezervatu. Ali mi rezervat ne prihvaćamo jer to
nije sloboda. Antifašisti u Hrvatskoj ne mogu biti poput Indijanaca svrstani u rezervate
koje će država štiti od okolnih fašista.
“…the state protects us here as if we were in a reservation. But we do not accept the
reservation, because that is not freedom. The antifascists in Croatia can’t be put in
reservations, like Indians, only to be protected from the surrounding fascists.”
The concept rezervat (“reservation”) is found only in this instance of the corpus, but it is highly
emblematic of the Cultural Model represented by the speaker. The antifascists are conceptualized
as Indians, old-settlers, a minority contained in an enclosed space designated by the Croatian state,
while the fascists are threating to extinguish their presence even from this small secure habitat, or
in this case, the memory of the antifascist uprising from the national cultural memory. The
metaphorical activation of the RESERVATION domain is an excellent way of mobilizing emotions,
reinforcing identity and moral values, and even vividly representing the repercussion of not
standing up to the political fight for existence that they are facing. However, this feeling also
perhaps contributes to the absence of the otherwise very frequent attribution of the Croatian state
as domovina (“homeland”) in speeches of the representatives of the Serb National Council in the
FRAMNAT corpus.
Identifying the speaker communities via the distribution of concepts
5
www.framnat.eu.

19. 81
Community identification methodology can be used for discerning the Cultural Model of
conceptualization related to a particular speaker (Illustration 6).
Figure 3.6 Graph of the relations between the 3,370 noun lemmas expressed by the sixty-four
speakers. The size of the labels corresponds to the overall frequency of the nouns connected with
the speaker.
By applying the algorithm for unfolding communities in the network (Blondel et al., 2008),
represented in Illustration 6, we can distinguish between the ten communities. The communities
of the speakers according to the similarity of the nouns they used in their speeches are shown in
Illustration 7 and in Table 10.

20. 82
Figure 3.7 The distribution of the speakers according to the communities organized by the common
use of the noun lexical concepts. The closeness of the nodes visualizes the similarity of the usage
of lexical concepts. The size of the labels corresponds to the degree of the connections.
Table 3.10 Communities of the speakers clustered according to the similarity of the nouns used in
their speeches.
Community Speakers % of the network
activation
1 Josip Bozanić, Franjo Komarica, Vjekoslav Huzjak,
Nikola Kekić, Juraj Jezerinac, student 6, student 5,
student 7, student 9, student 4
17,6 %
2 Franjo Habulin, Milorad Pupovac, Stjepan Mesić, Zoran
Pusić, Milan Surla, Milan Tankosić, Ivanka Roksandić,
Dragan Markovina
18,9 %
3 Mate Uzinić, Želimir Puljić, Đuro Hranić, Ivica Jagodić,
Manda Patko, student 3, student 1, student 2
12,6 %

21. 83
4 Bruna Esih, Dragan Čović, Ante Kutleša, Željko Reiner,
Idriz ef. Bešić, Zlatko Ževrnja, Borjana Krišto, Aziz ef.
Hasanović, Orest Wilczynski, student 8
11,7 %
5 Kolinda Grabar-Kitarović, Ivo Josipović, Josip Leko,
Milan Bandić, Dražimir Jukić, Ivan Vukić, Dubravka
Jurlina Alibegović, Margareta Mađerić, Madij Ismailov
11,3 %
6 Kristina Ikić Baniček, Nikola Budija 5, 1 %
7 Zoran Milanović, Tomislav Sopta, Maciej Szymanski,
Branko Lustig
8,2 %
8 Boris Prebeg, Frano Čirko 3,7 %
9 Ivica Glavota, Josipa Rimac 3,4 %
10 Nataša Mataušić, Nevenka Marinković, Imam Admir
Muhić, Ivica Vukelić
7,5 %
The percentage of the activation indicates the amount of lexical diversity of the particular cluster.
The higher the number, the greater the lexical diversity. By comparing ten communities in Table
10 with the distribution of speakers across the commemorations (Illustration 1), we can conclude
that similarities in the conceptualizations have strong correlations with the institutional affiliation
as well as with the particular commemoration. For instance, communities 1 and 3 contain mostly
Catholic Church representatives, but community 3 is focused more on the conceptualizations
specific to the Vukovar commemoration. Community 2 features the institutions and speakers that
promote the antifascist Cultural Model. Community 5 includes high-ranking active politicians,
with the exception of the former Prime Minister Milanović who is, according to the clustering
algorithm, conceptually related to community 7.
Subgraph of the Nouns-by-Speaker graph
By filtering the graph into subgraph communities (Illustrations 8-12) we can identify the key
common concepts, represented in the central region of the graph, in conjunction with the lexical
particularities related to the speaker, visualized on the margins.
Starting with the most lexically diverse community, community 2, we can identify that the salient
common lexical concepts are represented as the central nodes of the graph: jednakost (“equality”),
ustanak (“uprising”), borba (“struggle”), fašizam (“fascism”), antifašizam (“antifascism”), and
drug (“comrade”). On the edges of the graph are speaker specific concepts, such as volja (“will”)
for Stjepan Mesić, or falsificiranje (“falsification”) for Zoran Pusić.

22. 84
Figure 3.8 Graph of the community created by the common use of the 649 noun lexical concepts
expressed by ten speakers: Franjo Habulin, Milorad Pupovac, Stjepan Mesić, Zoran Pusić, Milan
Surla, Milan Tankosić, Ivanka Roksandić, Dragan Markovina. The size of the nodes corresponds
to the degree of the connections.
Table 3.11 List of salient commonly used noun lexical concepts in the Pupovac et al., community.
Lexeme (hr) Translation In-degree Frequency
borba struggle 9 76
dan day 9 67
kraj end 9 39
borac fighter 8 47
ustanak uprising 8 67
fašizam fascism 7 51
antifašizam antifascism 7 44
vrijednost value 7 29
drug comrade 7 18

23. 85
Srbin Serb 6 35
činjenica fact 6 28
ideja idea 6 33
Europa Europe 6 24
Srb Srb 6 29
prošlost history 6 16
početak beginning 6 12
čast honour 6 9
većina majority 6 11
mjesto place 5 40
vlast government 5 22
okupator occupator 5 21
pokret movement 5 20
otpor resistance 5 12
demokracija democracy 5 10
slava glory 5 12
jednakost equality 5 14
mjesec month 5 9
razdoblje period 5 6
ustaša Ustaša 4 16
Jugoslavija Yugoslavia 4 26
vjera faith 4 19
drugarica comrade 4 15
prostor space 4 13
odred unit 4 14
nacija nation 4 12
zakon law 4 12

24. 86
laž lie 4 12
partija party 4 13
bratstvo brotherhood 4 13
obzir consideration 4 10
lipanj June 4 9
skup gathering 4 8
uzvanik guest 4 7
Rom Gypsy 4 9
srpanj July 4 10
tekovina heritage 4 6
temelj foundation 4 9
smisao sense 4 7
prilika chance 4 7
revizionizam revisionism 4 7
negiranje negation 4 6
The graph representation of the concepts in community 1 (Illustration 9) mostly comprised of
Catholic bishops shows the prevalence of the common theological models related to the nouns
such as Bog (“God”), grijeh (“sin”), vjernik (“believer”), etc., as well as the conceptualization
peculiarities of different clerical speakers.

25. 87
Figure 3.9 Graph of the community created by the common use of the 601 noun lexical concepts
expressed by eleven speakers: Franjo Komarica, Josip Bozanić, Vjekoslav Huzjak, Nikola Kekić,
and students. The size of the nodes corresponds to the degree of the connections.
Table 3.12 List of salient common used noun lexical concepts in the Bozanić et al., community
Lexeme (hr) Translation In-degree Frequency
mir peace 9 39
čovjek man 8 118
život life 7 89
zemlja land 7 35
svijet world 7 21
smrt death 6 29
vrijeme time 6 24
sestra sister 5 41

26. 88
braća brother 5 38
vjernik believer 5 29
povijest history 5 28
nada hope 5 26
bog god 5 41
svjetlo light 5 13
Krist Christ 5 15
duša soul 5 13
evanđelje gospel 5 9
san dream 5 5
zajedništvo unity 4 17
kršćanin Christianity 4 15
Bog god 4 24
snaga strength 4 13
grijeh sin 4 21
dostojanstvo dignity 4 10
Community 3 is also comprised mostly of clerical speakers related to the Vukovar
commemoration. The central common concepts specific for this cluster are: Vukovar (“Vukovar”),
grad (“city”), groblje (“cemetery”), bol (“pain”), ponos (“pride”), ljubav (“love”), but also the
concepts that frame the catholic Cultural model: oltar (“altar”), oltar domovine (“altar of
homeland”), pijetet (“piety”), and nadbiskup (“archbishop”).

27. 89
Figure 3.10 Graph of the community created by the common use of the 434 noun lexical concepts
expressed by Mate Uzinić, Želimir Puljić, Đuro Hranić, Manda Patko. The size of the nodes
corresponds to the degree of the connections.
Table 3.13 List of most common used noun lexical concepts in the Uzinić et al., community.
Lexeme (hr) Translation In-degree Frequency
grad city 8 70
Vukovar Vukovar 8 69
srce heart 6 22
ljubav love 6 14
majka mother 6 10
dio part 6 13
riječ word 5 22
ponos pride 5 13
monsinjor monsignor 4 17

28. 90
pijetet piety 4 17
način method 4 11
molitva prayer 4 10
groblje cemetery 4 10
sin son 4 8
nadbiskup archbishop 4 10
Ivan Ivan 4 8
oltar altar 4 4
bol pain 4 5
stradanje suffering 3 18
Isus Jesus 3 12
biskup bishop 3 9
predstavnik representative 3 7
slavlje celebration 3 8
Kolona row 3 7
Žena woman 3 7
Čitanje reading 3 7
Agresija aggression 3 5
Studeni November 3 6
Duh spirit 3 6
Grob grave 3 4
Tuga sadness 3 3
Rana wound 3 4
spasenje salvation 3 4
Penava Penava 3 3

29. 91
The community represented by the speakers Bruna Esih, Dragan Čović, Ante Kutleša, Željko
Reiner, and others (Illustration 11) is related to the Bleiburg commemoration. The common
concepts specific to this community activate the victim-crime frame (Table 14).
Figure 3.11 Graph of the community created by the common use of the 401 noun lexical concepts
expressed by Bruna Esih, Dragan Čović, Ante Kutleša, Željko Reiner, Idriz ef. Bešić, Zlatko
Ževrnja, Borjana Krišto, Aziz ef. Hasanović, Orest Wilczynski, student 8. The size of the nodes
corresponds to the degree of the connections.
Table 3.14 List of most common used noun lexical concepts in the Esih et al., community.
Lexeme (hr) Translation In-degree Frequency
žrtva victim 9 81
zločin crime 8 58
put path 8 40
stratište execution site 8 14

30. 92
broj number 7 18
Bleiburg Bleiburg 6 26
komemoracija commemoration 6 19
strana side 6 13
tragedija tragedy 6 18
dijete child 5 14
prijatelj friend 5 13
simbol symbol 5 11
jama pit 4 10
znak sign 4 7
dolazak arrival 4 9
kost bone 4 7
milost mercy 4 6
počivalište resting place 4 5
pripadnost affiliation 4 4
Community 5 is comprised of nine speakers, among which are high ranking political officials:
Kolinda Grabar Kitarović, Ivo Josipović, Josip Leko, Milan Bandić. The list of lemmas (table 15)
shows that the most common concepts in this community frame the typical narrative of the
homeland rat ‘war’ for sloboda ‘freedom’, fought by branitelj ‘defenders’, as well as the political
frames of budućnost “future” and napredak “progress”.

31. 93
Figure 3.12 Graph of the community created by the common use of 384 noun lexical concepts
expressed by Kolinda Grabar-Kitarović, Ivo Josipović, Josip Leko, Milan Bandić, Dražimir Jukić,
Ivan Vukić, Dubravka Jurlina Alibegović, Margareta Mađerić, Madij Ismailov. The size of the
nodes corresponds to the degree of the connections.
Table 3.15 List of the most commonly used noun lexical concepts
Lexeme (hr) Translation In-degree Frequency
rat war 9 67
sloboda freedom 9 54
Hrvatska Croatia 8 146
poštovanje respect 7 13
branitelj defender 6 54
ime name 6 21
hvala gratitude 6 13
gospođa madam 6 13

32. 94
obitelj family 6 15
sabor parliament 6 12
zlo evil 5 30
građanin citizen 5 20
društvo society 5 15
gospoda madam 5 14
Tuđman Tuđman 5 12
cilj goal 5 11
članica member 5 7
oluja storm 4 40
budućnost future 4 19
napredak progress 4 8
pripadnik member 4 8
proslava celebration 4 8
ideal ideal 4 6
hrabrost courage 4 5
domoljub patriot 4 6
solidarnost solidarity 4 6
gradonačelnica mayor 4 4
odlučnost determination 4 4
The identified salient lexical concepts frame the theme of the commemoration and play mayor role
in the conceptualization process.
Extracting the related concepts with the coordinated construction
By extending the semantical-syntactical relations between tokens on the level of lemmas, it is
possible to summarize the patterns of conceptualization for any individual text or the whole

33. 95
FRAMNAT 2014-2016 corpus of commemoration speeches. This can provide means to analyze
the syntactic dependencies as a measure of cognitive focus and cognitive entrenchment. In this
section I will present the ontological analysis of the concepts in the FRAMNAT database extracted
by the coordinated construction, as well as the analysis of the direct object construction.
The coordinated construction is a set of collocated words that co-occur connected by conjunctions
and (sometimes or). The idea is that people in discourse frequently use conjunction and connect
ontologically similar classes of entities (Perak, 2017). For instance, the most frequent collocated
words in coordinated construction are sister and brother (40 occurrences). This is the most frequent
coordinated collocation in other corpuses also, such as English 13 Giga word corpus enTenTen
with 138,239 matches,6
and Croatian 1,4 Giga words corpus hrwac22 0.70 with 9,388 hits.7
In
the FRAMNAT 2014-2016 corpus the coordinated construction is retrieved by using the
conjunction dependency ‘conj’. 8
By creating a conceptual network using all coordinated
collocations between nouns and assigning the frequency of the occurrence as the weight of the
Force layout graph we can extract the ontological network of the conceptual entities activated by
the FRAMNAT speeches. The conceptual network is presented in Illustration 13 with frequencies
of the collocations expressed on the relationships label.
6
https://the.sketchengine.co.uk/corpus/wsketch?corpname=preloaded%2Fententen13_tt2_1&reload=&lemma=sister
&lpos=-
n&usesubcorp=&minfreq=auto&minscore=0.0&maxitems=25&sort_ws_columns=s&show_lemma_coverage=0&sh
ow_lcm=0&show_lcm=1&clustercolls=0&minsim=0.15&structured=0&structured=1&min_unary_score=0.0&min_
mwlink_freq=100&nr_ws_cols=5&bim_lang=
7
https://the.sketchengine.co.uk/corpus/wsketch?corpname=preloaded%2Fhrwac22&reload=&lemma=sestra&lpos=-
n&usesubcorp=&minfreq=auto&minscore=0.0&maxitems=25&sort_ws_columns=s&show_lemma_coverage=0&sh
ow_lcm=0&show_lcm=1&clustercolls=0&minsim=0.15&structured=0&structured=1&min_unary_score=0.0&min_
mwlink_freq=100&nr_ws_cols=5&bim_lang=
8
http://universaldependencies.org/u/dep/conj.html

34. 96
Figure 3.13 The network of 1,677 noun lemmas in the FRAMNAT 2014-2016 corpus connected
by 2,335 coordinated collocations. The size of the node and label is determined by the amount of
the interconnectedness with other words. The layout is organized by Force Atlas algorithm.9
The graph (Illustration 13) represents 1,677 entities conceptualized by various texts in the
FRAMNAT 2014-2016 corpus that are connected with some other entity via the conjunction i
“and.” The size of the node and label is larger depending on the number of connections with other
words. If we filter out the nodes that have five or more connections we get 235 nodes (14%) with
830 connections (35.6%) (Illustration 14).
9
The interactive network can be found at the web page framnat.eu.

35. 97
Figure 3.14 The network of 234 nodes (noun lemmas) and 826 edges (relations) with five or more
coordinated relations in the FRAMNAT 2014-2016 corpus. The size of a node and label is
determined by the amount of the interconnectedness with other words. The layout is organized by
ForceAtlas algorithm.
The graph (Illustration 14) of lemmas with five or more connections with other lemmas visualizes
the conceptually salient and ontologically related concepts in the FRAMNAT 2014-2016 corpus.
The central nodes of the graph are the concepts related to the words žrtva (“victim”) and narod
(“people”), and Hrvatska (“Croatia”) that conceptualize the central concepts of the
commemorative speeches. In the upper region of the graph are two communities of concepts
related to the group identities such as prijatelj (“friend”), čovjek (“man”), predstavnik
(“representative”), antifašist (“antifascist”), and branitelj (“defender”). Concepts represented in
the lower central region, like rat (“war”), smrt (“death”), nasilje (“violence”) conceptualize the
violent nature of the commemorated events. In the lower part of the graph are concepts like ljubav
(“love”), mir (“peace”), istina (“truth”), and pravda (“justice”) that convey the psychological states
and the socially desirable modes of interactions along with their ontologically related opposites
such as bol (“pain”), strah (“fear”), and ustaštvo (“Ustaša-ism).

36. 98
Conclusion
This chapter deals with the analysis of the commemoration practices from the perspective of the
public communication acts that construe the networks of culturally distributed cognition and
conceptualizations. We have shown how this process evolves around speakers that act as the agents
of immediate conceptual and gradual cultural dissemination. The content of their message is
framed by the salient concepts from a cultural model, or the worldview, that speakers share by
institutional affiliation. The corpus analysis measured the frequency of the activated concepts in
speeches by speakers and institutions. By using the graph theory algorithms on the level of lexical
concepts we classified sixty-four speakers and eighteen supporting institutions according to the
3,370 invoked noun concepts at seven commemorations. The classification process has revealed
distinct communities of speakers and their shared choice of salient concepts and strategies of
framing the affective dispositions and cognitive processes that form the basis for the construction
of group identities, interaction and communication practices, political agenda and dominant
cultural model of national identity in general.
The institution affiliation is by no means irrelevant for the effectiveness of the speaker’s cultural
dissemination, as the speaker implicitly projects a prototypical institutional Cultural model while
tacitly activating a small number of highly potent conceptual frames. For instance, the phrase oltar
domovine ‘the altar of the homeland’ (Table 13) is used to frame the death of people in war as a
žrtva ‘sacrifice.’ The frame activates the conceptual mapping of “necessity,” “obedience,”
“usefulness,” and the “divine” features of religious ritual as described in the Bible myths onto the
features of those killed. It neutralizes the immediate psychological negative effect, fostering a
pragmatic implication of usefulness of their sacrifice that solidifies the socially constructed group
identity in terms of the sacred Christian narrative. However, this framing can be contra productive
for speakers affiliated with the institutions that oppose that particular belief system. Therefore,
they activate a different set of frames, such as borci “fighters,” ustanak “uprising” (Table 11), with
different pragmatic inferences that effectively produce the same social functions of social
cohesion. This process can be called contested conceptual framing. We have identified the salient
contested frames in the lists of commonly used noun lexical concepts for different speaker
communities (Tables 11-15) that generally correlate with the political identity and the type of
commemoration.
Furthermore, the study shows how the institution extends not just the psychological, but the
political and economic power to the individual representative by the practical organization of their
presence on the commemorations, harnessing in turn the effects of the conceptual dissemination,
promoted political agenda and ideology. The indication of this hegemonic process can be measured
in terms of speech length (Table 5, 6) as well as the frequency of concepts associated with speakers
(Illustration 6, 7) and institutions (Illustration 5). The results of our analysis clearly show the
specter of institutional coverage on different commemorations (Illustration 1) and a dominance of
Catholic institutions in terms of the framing and conceptualization of the national model of

37. 99
commemoration practices and their pragmatic implications for the process of construction of
identity.
Bibliography
Bergen, B., Chang, N., & Narayan, S. 2004. Simulated action in an embodied construction
grammar. In Proceedings of the Annual Meeting of the Cognitive Science Society (Vol. 26, No.
26).
Blondel, V.D., Guillaume, J-L., Lambiotte, R. and Lefebvre, E. 2008. Fast unfolding of
communities in large networks. Journal of Statistical Mechanics: Theory and Experiment 10.
Charteris-Black, J. 2005. Politicians and Rhetoric. New York: Palgrave Macmillan.
Charteris-Black, J. 2006. The Communication of Leadership. The Design of Leadership Style.
New York: Routledge.
D'Andrade, R. 1987. A folk model of the mind. In: Holland, D., & Quinn, N., eds. Cultural
models in language and thought. Cambridge University Press.
Hogan, P. C. 2009. Understanding Nationalism: On Narrative, Cognitive Science, and Identity.
Columbus: The Ohio State University Press.
Lakoff, G. and Johnson, M. 1999. Philosophy in the Flesh: The Embodied Mind and Its
Challenge to Western Thought. New York: Basic Books.
Lakoff, G. 1987. Women, Fire, and Dangerous Things: What Categories Reveal About the Mind.
Chicago: University of Chicago.
Lakoff, G. 2008. The Neural Theory of Metaphor. In: Gibbs, R., ed. The Cambridge Handbook
of Metaphor and Thought. New York: Cambridge University Press.
Ljubešić, N., Klubička, F., Agić, Ž., and Jazbec, I. P. 2016. New inflectional lexicons and
training corpora for improved morphosyntactic annotation of Croatian and Serbian. In: Tenth
International Conference on Language Resources and Evaluation (LREC 2016).
Kövecses, Z. 2000. Metaphor and Emotion: Language, Culture, and Body in Human Feeling.
New York: Cambridge.
Ma’iz, R. 2003. Framing the Nation: three Rival Versions of Contemporary Nationalist Ideology.
Journal of Political Ideologies 8 (3): 251-267.
Pavlaković, V. and Perak, B. 2017. How Does This Monument Make You Feel? Measuring
Emotional Responses to War Memorials in Croatia. In: Sindbæk Andersen, T. and Törnquist-

38. 100
Plewa, B., eds. The Twentieth Century in European Memory: Transcultural Mediation and
Reception. Leiden: Brill.
Perak, B. 2017. Pojmovne mreže leksema neugodnih i nepobuđenih emocija. Usporedba
pojmova tuga, jad, žalost, depresija i sjeta u koordiniranoj konstrukciji. Usporedba pojmova
tuga, jad, žalost, depresija i sjeta u koordiniranoj konstrukciji. In Stolac, D. and Vlastelić, A. eds.
Jezik kao predmet proučavanja i jezik kao predmet poučavanja. Zbornik radova s
međunarodnoga znanstvenog skupa Hrvatskoga društva za primijenjenu lingvistiku održanoga
od 12. do 14. svibnja 2016. godine u Rijeci. Zagreb: Srednja Europa.
Petričušić, A., Čehulić, M. and Čepo, D. 2017. Gaining Political Power by Utilizing Opportunity
Structures: An Analysis of the Conservative Religious-Political Movement in Croatia. Politička
misao: časopis za politologiju 54(4): 61-84. Retrieved from https://hrcak.srce.hr/190338
Searle, J. 2010. Making the social world: The structure of human civilization. New York: Oxford
University Press.